A propensity-matched analysis of real-world outcomes comparing direct oral anticoagulants and warfarin in patients diagnosed with portal vein thrombosis

Abstract

Abstract Introduction: The management of portal vein thrombosis (PVT) usually includes anticoagulation therapy, with frequently used anticoagulants such as warfarin and direct oral anticoagulants (DOACs). In this real-world study, we compare clinical outcomes in PVT patients who received warfarin and DOACs. Methods: Patients ≥18 years old with PVT between January 2015 and December 2020 were identified from 105 health care organizations in the TriNetX research network. Cohorts consisted of patients who received warfarin (n=7,497) or DOACs (n=9,995). Propensity score matching yielded well-balanced cohorts (5,905 patients each). Outcomes evaluated during a 180-day follow-up period were gastrointestinal (GI) bleeding, death, hepatic decompensation (ascites, esophageal varices with bleeding, hepatic encephalopathy), and other bleeding (retroperitoneal hematoma, intracerebral hemorrhage). Results: Warfarin-treated patients had comparable GI bleeding risk with DOACs (7.3% vs. 6.4%; risk ratio [RR]: 1.135, 95% CI: 0.993–1.296; p=0.063). Warfarin was also related to substantially reduced mortality (11.5% vs. 14.3%; RR: 0.802, 95% CI: 0.730–0.882; p<0.001) but increased risks for hepatic decompensation (25.3% vs. 21.9%; RR: 1.153, 95% CI: 1.080–1.231; p<0.001) and other bleeding complications (1.0% vs. 0.6%; RR: 1.758, 95% CI: 1.148–2.691; p=0.009). Kaplan-Meier analyses reinforced these results, showing significantly better survival with warfarin therapy (log-rank p<0.001) but with increased risk for hepatic decompensation (p<0.001) and for other bleeding complications (p=0.010). Conclusion: In PVT patients, warfarin therapy was linked to reduced short-term mortality but increased rates of hepatic decompensation and other major bleeding in comparison to DOACs. The results highlight the necessity for risk-benefit based individualized therapeutic choices in PVT patients.